+/*
+ * If the string contains escape sequences,
+ * replace it with an allocated rendering and return 1,
+ * such that the caller can free it after use.
+ * Otherwise, do nothing and return 0.
+ */
+static int
+render_string(char **public, size_t *psz)
+{
+ const char *src, *scp, *addcp, *seq;
+ char *dst;
+ size_t ssz, dsz, addsz;
+ char utfbuf[7], res[6];
+ int seqlen, unicode;
+
+ res[0] = '\\';
+ res[1] = '\t';
+ res[2] = ASCII_NBRSP;
+ res[3] = ASCII_HYPH;
+ res[4] = ASCII_BREAK;
+ res[5] = '\0';
+
+ src = scp = *public;
+ ssz = *psz;
+ dst = NULL;
+ dsz = 0;
+
+ while (scp < src + *psz) {
+
+ /* Leave normal characters unchanged. */
+
+ if (strchr(res, *scp) == NULL) {
+ if (dst != NULL)
+ dst[dsz++] = *scp;
+ scp++;
+ continue;
+ }
+
+ /*
+ * Found something that requires replacing,
+ * make sure we have a destination buffer.
+ */
+
+ if (dst == NULL) {
+ dst = mandoc_malloc(ssz + 1);
+ dsz = scp - src;
+ memcpy(dst, src, dsz);
+ }
+
+ /* Handle single-char special characters. */
+
+ switch (*scp) {
+ case '\\':
+ break;
+ case '\t':
+ /* FALLTHROUGH */
+ case ASCII_NBRSP:
+ dst[dsz++] = ' ';
+ scp++;
+ continue;
+ case ASCII_HYPH:
+ dst[dsz++] = '-';
+ /* FALLTHROUGH */
+ case ASCII_BREAK:
+ scp++;
+ continue;
+ default:
+ abort();
+ }
+
+ /*
+ * Found an escape sequence.
+ * Read past the slash, then parse it.
+ * Ignore everything except characters.
+ */
+
+ scp++;
+ if (mandoc_escape(&scp, &seq, &seqlen) != ESCAPE_SPECIAL)
+ continue;
+
+ /*
+ * Render the special character
+ * as either UTF-8 or ASCII.
+ */
+
+ if (write_utf8) {
+ unicode = mchars_spec2cp(mchars, seq, seqlen);
+ if (unicode <= 0)
+ continue;
+ addsz = utf8(unicode, utfbuf);
+ if (addsz == 0)
+ continue;
+ addcp = utfbuf;
+ } else {
+ addcp = mchars_spec2str(mchars, seq, seqlen, &addsz);
+ if (addcp == NULL)
+ continue;
+ if (*addcp == ASCII_NBRSP) {
+ addcp = " ";
+ addsz = 1;
+ }
+ }
+
+ /* Copy the rendered glyph into the stream. */
+
+ ssz += addsz;
+ dst = mandoc_realloc(dst, ssz + 1);
+ memcpy(dst + dsz, addcp, addsz);
+ dsz += addsz;
+ }
+ if (dst != NULL) {
+ *public = dst;
+ *psz = dsz;
+ }
+
+ /* Trim trailing whitespace and NUL-terminate. */
+
+ while (*psz > 0 && (*public)[*psz - 1] == ' ')
+ --*psz;
+ if (dst != NULL) {
+ (*public)[*psz] = '\0';
+ return(1);
+ } else
+ return(0);
+}
+
+static void
+dbadd_mlink(const struct mlink *mlink)
+{
+ size_t i;
+
+ i = 1;
+ SQL_BIND_TEXT(stmts[STMT_INSERT_LINK], i, mlink->dsec);
+ SQL_BIND_TEXT(stmts[STMT_INSERT_LINK], i, mlink->arch);
+ SQL_BIND_TEXT(stmts[STMT_INSERT_LINK], i, mlink->name);
+ SQL_BIND_INT64(stmts[STMT_INSERT_LINK], i, mlink->mpage->pageid);
+ SQL_STEP(stmts[STMT_INSERT_LINK]);
+ sqlite3_reset(stmts[STMT_INSERT_LINK]);
+}
+
+static void
+dbadd_mlink_name(const struct mlink *mlink)
+{
+ uint64_t bits;
+ size_t i;
+
+ dbadd_mlink(mlink);
+
+ i = 1;
+ SQL_BIND_INT64(stmts[STMT_SELECT_NAME], i, mlink->mpage->pageid);
+ bits = NAME_FILE & NAME_MASK;
+ if (sqlite3_step(stmts[STMT_SELECT_NAME]) == SQLITE_ROW) {
+ bits |= sqlite3_column_int64(stmts[STMT_SELECT_NAME], 0);
+ sqlite3_reset(stmts[STMT_SELECT_NAME]);
+ }
+
+ i = 1;
+ SQL_BIND_INT64(stmts[STMT_INSERT_NAME], i, bits);
+ SQL_BIND_TEXT(stmts[STMT_INSERT_NAME], i, mlink->name);
+ SQL_BIND_INT64(stmts[STMT_INSERT_NAME], i, mlink->mpage->pageid);
+ SQL_STEP(stmts[STMT_INSERT_NAME]);
+ sqlite3_reset(stmts[STMT_INSERT_NAME]);
+}
+
+/*
+ * Flush the current page's terms (and their bits) into the database.
+ * Wrap the entire set of additions in a transaction to make sqlite be a
+ * little faster.
+ * Also, handle escape sequences at the last possible moment.
+ */
+static void
+dbadd(struct mpage *mpage)
+{
+ struct mlink *mlink;
+ struct str *key;
+ char *cp;
+ size_t i;
+ unsigned int slot;
+ int mustfree;
+
+ mlink = mpage->mlinks;
+
+ if (nodb) {
+ for (key = ohash_first(&names, &slot); NULL != key;
+ key = ohash_next(&names, &slot))
+ free(key);
+ for (key = ohash_first(&strings, &slot); NULL != key;
+ key = ohash_next(&strings, &slot))
+ free(key);
+ if (0 == debug)
+ return;
+ while (NULL != mlink) {
+ fputs(mlink->name, stdout);
+ if (NULL == mlink->next ||
+ strcmp(mlink->dsec, mlink->next->dsec) ||
+ strcmp(mlink->fsec, mlink->next->fsec) ||
+ strcmp(mlink->arch, mlink->next->arch)) {
+ putchar('(');
+ if ('\0' == *mlink->dsec)
+ fputs(mlink->fsec, stdout);
+ else
+ fputs(mlink->dsec, stdout);
+ if ('\0' != *mlink->arch)
+ printf("/%s", mlink->arch);
+ putchar(')');
+ }
+ mlink = mlink->next;
+ if (NULL != mlink)
+ fputs(", ", stdout);
+ }
+ printf(" - %s\n", mpage->desc);
+ return;
+ }
+
+ if (debug)
+ say(mlink->file, "Adding to database");
+
+ cp = mpage->desc;
+ i = strlen(cp);
+ mustfree = render_string(&cp, &i);
+ i = 1;
+ SQL_BIND_TEXT(stmts[STMT_INSERT_PAGE], i, cp);
+ SQL_BIND_INT(stmts[STMT_INSERT_PAGE], i, mpage->form);
+ SQL_STEP(stmts[STMT_INSERT_PAGE]);
+ mpage->pageid = sqlite3_last_insert_rowid(db);
+ sqlite3_reset(stmts[STMT_INSERT_PAGE]);
+ if (mustfree)
+ free(cp);
+
+ while (NULL != mlink) {
+ dbadd_mlink(mlink);
+ mlink = mlink->next;
+ }
+ mlink = mpage->mlinks;
+
+ for (key = ohash_first(&names, &slot); NULL != key;
+ key = ohash_next(&names, &slot)) {
+ assert(key->mpage == mpage);
+ i = 1;
+ SQL_BIND_INT64(stmts[STMT_INSERT_NAME], i, key->mask);
+ SQL_BIND_TEXT(stmts[STMT_INSERT_NAME], i, key->key);
+ SQL_BIND_INT64(stmts[STMT_INSERT_NAME], i, mpage->pageid);
+ SQL_STEP(stmts[STMT_INSERT_NAME]);
+ sqlite3_reset(stmts[STMT_INSERT_NAME]);
+ free(key);
+ }
+ for (key = ohash_first(&strings, &slot); NULL != key;
+ key = ohash_next(&strings, &slot)) {
+ assert(key->mpage == mpage);
+ i = 1;
+ SQL_BIND_INT64(stmts[STMT_INSERT_KEY], i, key->mask);
+ SQL_BIND_TEXT(stmts[STMT_INSERT_KEY], i, key->key);
+ SQL_BIND_INT64(stmts[STMT_INSERT_KEY], i, mpage->pageid);
+ SQL_STEP(stmts[STMT_INSERT_KEY]);
+ sqlite3_reset(stmts[STMT_INSERT_KEY]);
+ free(key);
+ }
+}
+
+static void
+dbprune(void)
+{
+ struct mpage *mpage;
+ struct mlink *mlink;
+ size_t i;
+ unsigned int slot;
+
+ if (0 == nodb)
+ SQL_EXEC("BEGIN TRANSACTION");
+
+ for (mpage = ohash_first(&mpages, &slot); NULL != mpage;
+ mpage = ohash_next(&mpages, &slot)) {
+ mlink = mpage->mlinks;
+ if (debug)
+ say(mlink->file, "Deleting from database");
+ if (nodb)
+ continue;
+ for ( ; NULL != mlink; mlink = mlink->next) {
+ i = 1;
+ SQL_BIND_TEXT(stmts[STMT_DELETE_PAGE],
+ i, mlink->dsec);
+ SQL_BIND_TEXT(stmts[STMT_DELETE_PAGE],
+ i, mlink->arch);
+ SQL_BIND_TEXT(stmts[STMT_DELETE_PAGE],
+ i, mlink->name);
+ SQL_STEP(stmts[STMT_DELETE_PAGE]);
+ sqlite3_reset(stmts[STMT_DELETE_PAGE]);
+ }
+ }
+
+ if (0 == nodb)
+ SQL_EXEC("END TRANSACTION");
+}
+
+/*
+ * Close an existing database and its prepared statements.
+ * If "real" is not set, rename the temporary file into the real one.
+ */
+static void
+dbclose(int real)
+{
+ size_t i;
+ int status;
+ pid_t child;
+
+ if (nodb)
+ return;
+
+ for (i = 0; i < STMT__MAX; i++) {
+ sqlite3_finalize(stmts[i]);
+ stmts[i] = NULL;
+ }
+
+ sqlite3_close(db);
+ db = NULL;
+
+ if (real)
+ return;
+
+ if ('\0' == *tempfilename) {
+ if (-1 == rename(MANDOC_DB "~", MANDOC_DB)) {
+ exitcode = (int)MANDOCLEVEL_SYSERR;
+ say(MANDOC_DB, "&rename");
+ }
+ return;
+ }
+
+ switch (child = fork()) {
+ case -1:
+ exitcode = (int)MANDOCLEVEL_SYSERR;
+ say("", "&fork cmp");
+ return;
+ case 0:
+ execlp("cmp", "cmp", "-s",
+ tempfilename, MANDOC_DB, NULL);
+ say("", "&exec cmp");
+ exit(0);
+ default:
+ break;
+ }
+ if (-1 == waitpid(child, &status, 0)) {
+ exitcode = (int)MANDOCLEVEL_SYSERR;
+ say("", "&wait cmp");
+ } else if (WIFSIGNALED(status)) {
+ exitcode = (int)MANDOCLEVEL_SYSERR;
+ say("", "cmp died from signal %d", WTERMSIG(status));
+ } else if (WEXITSTATUS(status)) {
+ exitcode = (int)MANDOCLEVEL_SYSERR;
+ say(MANDOC_DB,
+ "Data changed, but cannot replace database");
+ }
+
+ *strrchr(tempfilename, '/') = '\0';
+ switch (child = fork()) {
+ case -1:
+ exitcode = (int)MANDOCLEVEL_SYSERR;
+ say("", "&fork rm");
+ return;
+ case 0:
+ execlp("rm", "rm", "-rf", tempfilename, NULL);
+ say("", "&exec rm");
+ exit((int)MANDOCLEVEL_SYSERR);
+ default:
+ break;
+ }
+ if (-1 == waitpid(child, &status, 0)) {
+ exitcode = (int)MANDOCLEVEL_SYSERR;
+ say("", "&wait rm");
+ } else if (WIFSIGNALED(status) || WEXITSTATUS(status)) {
+ exitcode = (int)MANDOCLEVEL_SYSERR;
+ say("", "%s: Cannot remove temporary directory",
+ tempfilename);
+ }
+}
+
+/*
+ * This is straightforward stuff.
+ * Open a database connection to a "temporary" database, then open a set
+ * of prepared statements we'll use over and over again.
+ * If "real" is set, we use the existing database; if not, we truncate a
+ * temporary one.
+ * Must be matched by dbclose().
+ */
+static int
+dbopen(int real)
+{
+ const char *sql;
+ int rc, ofl;
+
+ if (nodb)
+ return(1);
+
+ *tempfilename = '\0';
+ ofl = SQLITE_OPEN_READWRITE;
+
+ if (real) {
+ rc = sqlite3_open_v2(MANDOC_DB, &db, ofl, NULL);
+ if (SQLITE_OK != rc) {
+ exitcode = (int)MANDOCLEVEL_SYSERR;
+ if (SQLITE_CANTOPEN != rc)
+ say(MANDOC_DB, "%s", sqlite3_errstr(rc));
+ return(0);
+ }
+ goto prepare_statements;
+ }
+
+ ofl |= SQLITE_OPEN_CREATE | SQLITE_OPEN_EXCLUSIVE;
+
+ remove(MANDOC_DB "~");
+ rc = sqlite3_open_v2(MANDOC_DB "~", &db, ofl, NULL);
+ if (SQLITE_OK == rc)
+ goto create_tables;
+ if (MPARSE_QUICK & mparse_options) {
+ exitcode = (int)MANDOCLEVEL_SYSERR;
+ say(MANDOC_DB "~", "%s", sqlite3_errstr(rc));
+ return(0);
+ }
+
+ (void)strlcpy(tempfilename, "/tmp/mandocdb.XXXXXX",
+ sizeof(tempfilename));
+ if (NULL == mkdtemp(tempfilename)) {
+ exitcode = (int)MANDOCLEVEL_SYSERR;
+ say("", "&%s", tempfilename);
+ return(0);
+ }
+ (void)strlcat(tempfilename, "/" MANDOC_DB,
+ sizeof(tempfilename));
+ rc = sqlite3_open_v2(tempfilename, &db, ofl, NULL);
+ if (SQLITE_OK != rc) {
+ exitcode = (int)MANDOCLEVEL_SYSERR;
+ say("", "%s: %s", tempfilename, sqlite3_errstr(rc));
+ return(0);
+ }
+
+create_tables:
+ sql = "CREATE TABLE \"mpages\" (\n"
+ " \"desc\" TEXT NOT NULL,\n"
+ " \"form\" INTEGER NOT NULL,\n"
+ " \"pageid\" INTEGER PRIMARY KEY AUTOINCREMENT NOT NULL\n"
+ ");\n"
+ "\n"
+ "CREATE TABLE \"mlinks\" (\n"
+ " \"sec\" TEXT NOT NULL,\n"
+ " \"arch\" TEXT NOT NULL,\n"
+ " \"name\" TEXT NOT NULL,\n"
+ " \"pageid\" INTEGER NOT NULL REFERENCES mpages(pageid) "
+ "ON DELETE CASCADE\n"
+ ");\n"
+ "CREATE INDEX mlinks_pageid_idx ON mlinks (pageid);\n"
+ "\n"
+ "CREATE TABLE \"names\" (\n"
+ " \"bits\" INTEGER NOT NULL,\n"
+ " \"name\" TEXT NOT NULL,\n"
+ " \"pageid\" INTEGER NOT NULL REFERENCES mpages(pageid) "
+ "ON DELETE CASCADE,\n"
+ " UNIQUE (\"name\", \"pageid\") ON CONFLICT REPLACE\n"
+ ");\n"
+ "\n"
+ "CREATE TABLE \"keys\" (\n"
+ " \"bits\" INTEGER NOT NULL,\n"
+ " \"key\" TEXT NOT NULL,\n"
+ " \"pageid\" INTEGER NOT NULL REFERENCES mpages(pageid) "
+ "ON DELETE CASCADE\n"
+ ");\n"
+ "CREATE INDEX keys_pageid_idx ON keys (pageid);\n";
+
+ if (SQLITE_OK != sqlite3_exec(db, sql, NULL, NULL, NULL)) {
+ exitcode = (int)MANDOCLEVEL_SYSERR;
+ say(MANDOC_DB, "%s", sqlite3_errmsg(db));
+ sqlite3_close(db);
+ return(0);
+ }
+
+prepare_statements:
+ if (SQLITE_OK != sqlite3_exec(db,
+ "PRAGMA foreign_keys = ON", NULL, NULL, NULL)) {
+ exitcode = (int)MANDOCLEVEL_SYSERR;
+ say(MANDOC_DB, "PRAGMA foreign_keys: %s",
+ sqlite3_errmsg(db));
+ sqlite3_close(db);
+ return(0);
+ }
+
+ sql = "DELETE FROM mpages WHERE pageid IN "
+ "(SELECT pageid FROM mlinks WHERE "
+ "sec=? AND arch=? AND name=?)";
+ sqlite3_prepare_v2(db, sql, -1, &stmts[STMT_DELETE_PAGE], NULL);
+ sql = "INSERT INTO mpages "
+ "(desc,form) VALUES (?,?)";
+ sqlite3_prepare_v2(db, sql, -1, &stmts[STMT_INSERT_PAGE], NULL);
+ sql = "INSERT INTO mlinks "
+ "(sec,arch,name,pageid) VALUES (?,?,?,?)";
+ sqlite3_prepare_v2(db, sql, -1, &stmts[STMT_INSERT_LINK], NULL);
+ sql = "SELECT bits FROM names where pageid = ?";
+ sqlite3_prepare_v2(db, sql, -1, &stmts[STMT_SELECT_NAME], NULL);
+ sql = "INSERT INTO names "
+ "(bits,name,pageid) VALUES (?,?,?)";
+ sqlite3_prepare_v2(db, sql, -1, &stmts[STMT_INSERT_NAME], NULL);
+ sql = "INSERT INTO keys "
+ "(bits,key,pageid) VALUES (?,?,?)";
+ sqlite3_prepare_v2(db, sql, -1, &stmts[STMT_INSERT_KEY], NULL);
+
+#ifndef __APPLE__
+ /*
+ * When opening a new database, we can turn off
+ * synchronous mode for much better performance.
+ */
+
+ if (real && SQLITE_OK != sqlite3_exec(db,
+ "PRAGMA synchronous = OFF", NULL, NULL, NULL)) {
+ exitcode = (int)MANDOCLEVEL_SYSERR;
+ say(MANDOC_DB, "PRAGMA synchronous: %s",
+ sqlite3_errmsg(db));
+ sqlite3_close(db);
+ return(0);
+ }
+#endif
+
+ return(1);
+}
+
+static void *
+hash_calloc(size_t n, size_t sz, void *arg)
+{
+
+ return(mandoc_calloc(n, sz));
+}
+
+static void *
+hash_alloc(size_t sz, void *arg)
+{
+
+ return(mandoc_malloc(sz));
+}
+
+static void
+hash_free(void *p, void *arg)
+{
+
+ free(p);
+}
+
+static int
+set_basedir(const char *targetdir, int report_baddir)
+{
+ static char startdir[PATH_MAX];
+ static int getcwd_status; /* 1 = ok, 2 = failure */
+ static int chdir_status; /* 1 = changed directory */
+ char *cp;
+
+ /*
+ * Remember the original working directory, if possible.
+ * This will be needed if the second or a later directory
+ * on the command line is given as a relative path.
+ * Do not error out if the current directory is not
+ * searchable: Maybe it won't be needed after all.
+ */
+ if (0 == getcwd_status) {
+ if (NULL == getcwd(startdir, sizeof(startdir))) {
+ getcwd_status = 2;
+ (void)strlcpy(startdir, strerror(errno),
+ sizeof(startdir));
+ } else
+ getcwd_status = 1;
+ }
+
+ /*
+ * We are leaving the old base directory.
+ * Do not use it any longer, not even for messages.
+ */
+ *basedir = '\0';
+
+ /*
+ * If and only if the directory was changed earlier and
+ * the next directory to process is given as a relative path,
+ * first go back, or bail out if that is impossible.
+ */
+ if (chdir_status && '/' != *targetdir) {
+ if (2 == getcwd_status) {
+ exitcode = (int)MANDOCLEVEL_SYSERR;
+ say("", "getcwd: %s", startdir);
+ return(0);
+ }
+ if (-1 == chdir(startdir)) {
+ exitcode = (int)MANDOCLEVEL_SYSERR;
+ say("", "&chdir %s", startdir);
+ return(0);
+ }
+ }
+
+ /*
+ * Always resolve basedir to the canonicalized absolute
+ * pathname and append a trailing slash, such that
+ * we can reliably check whether files are inside.
+ */
+ if (NULL == realpath(targetdir, basedir)) {
+ if (report_baddir || errno != ENOENT) {
+ exitcode = (int)MANDOCLEVEL_BADARG;
+ say("", "&%s: realpath", targetdir);
+ }
+ return(0);
+ } else if (-1 == chdir(basedir)) {
+ if (report_baddir || errno != ENOENT) {
+ exitcode = (int)MANDOCLEVEL_BADARG;
+ say("", "&chdir");
+ }
+ return(0);
+ }
+ chdir_status = 1;
+ cp = strchr(basedir, '\0');
+ if ('/' != cp[-1]) {
+ if (cp - basedir >= PATH_MAX - 1) {
+ exitcode = (int)MANDOCLEVEL_SYSERR;
+ say("", "Filename too long");
+ return(0);
+ }
+ *cp++ = '/';
+ *cp = '\0';
+ }
+ return(1);
+}
+
+static void
+say(const char *file, const char *format, ...)
+{
+ va_list ap;
+ int use_errno;
+
+ if ('\0' != *basedir)
+ fprintf(stderr, "%s", basedir);
+ if ('\0' != *basedir && '\0' != *file)
+ fputc('/', stderr);
+ if ('\0' != *file)
+ fprintf(stderr, "%s", file);
+
+ use_errno = 1;
+ if (NULL != format) {
+ switch (*format) {
+ case '&':
+ format++;
+ break;
+ case '\0':
+ format = NULL;
+ break;
+ default:
+ use_errno = 0;
+ break;
+ }
+ }
+ if (NULL != format) {
+ if ('\0' != *basedir || '\0' != *file)
+ fputs(": ", stderr);
+ va_start(ap, format);
+ vfprintf(stderr, format, ap);
+ va_end(ap);
+ }
+ if (use_errno) {
+ if ('\0' != *basedir || '\0' != *file || NULL != format)
+ fputs(": ", stderr);
+ perror(NULL);
+ } else
+ fputc('\n', stderr);
git.ckatri.com / mandoc.git / blobdiff